Major contact stability improvement

Laurent Coulon
Posts: 29
Joined: Tue Oct 05, 2010 9:36 pm

Re: Major contact stability improvement

Post by Laurent Coulon »

2.76
JHoule
Posts: 14
Joined: Tue May 22, 2012 9:30 pm

Re: Major contact stability improvement

Post by JHoule »

Indeed, it does make a smaller patch (see attachment).

Still a little tricky to get to work in our environment, though... :-(

I'm gonna try with the Bullet demos...
Attachments
contact.patch.gz
The previous 6 files, diff'd against Bullet 2.76 (proper baseline).
(18.95 KiB) Downloaded 645 times
JHoule
Posts: 14
Joined: Tue May 22, 2012 9:30 pm

Re: Major contact stability improvement

Post by JHoule »

Not much difference in the end: I still get pretty much the same errors.

I think I'll wait for a full package, if you can make one available...
Mako_energy02
Posts: 171
Joined: Sun Jan 17, 2010 4:47 am

Re: Major contact stability improvement

Post by Mako_energy02 »

I strongly encourage you to try and get these enhancements into the main Bullet repo if there are no licensing issues to prevent that. If you do want to add this to the main bullet distribution then the first step you have already done (this post). The second step is to add a report to the bullet issue tracker and label this as an enhancement, with a link pointing back to this thread. Then pray hard that Erwin notices this post.

You should include a basic description of what you did of course(in the report), and I believe there are no, or very high limits on what you can attach there(since it's google code).
Laurent Coulon
Posts: 29
Joined: Tue Oct 05, 2010 9:36 pm

Re: Major contact stability improvement

Post by Laurent Coulon »

Thanks,
I didn't know what the proper process was to do this. I will definitely try doing this.
User avatar
Erwin Coumans
Site Admin
Posts: 4221
Joined: Sun Jun 26, 2005 6:43 pm
Location: California, USA
Contact:

Re: Major contact stability improvement

Post by Erwin Coumans »

Thanks for sharing your ideas and work.

It will help if you update your work to the latest Bullet svn repository. I will create an up-to-date github repository, that makes it easier to fork/clone your modified version.
pico
Posts: 229
Joined: Sun Sep 30, 2007 7:58 am

Re: Major contact stability improvement

Post by pico »

Hi Laurent,

thanks for your work.

I would suggest to post such issues/patches here:
http://code.google.com/p/bullet/issues/list

So you can make sure Erwin actually notices them and integrates them into the main branch when possible.

(posted again in http://www.bulletphysics.org/Bullet/php ... 984#p27984)
Laurent Coulon
Posts: 29
Joined: Tue Oct 05, 2010 9:36 pm

Re: Major contact stability improvement

Post by Laurent Coulon »

Hi Erwin,
Thanks for the reply. I will try to make a patch for the latest bullet version. Let me know what the name of the repository is on github. Otherwise I can post it on Google code. Let me know.
VicariousEnt
Posts: 50
Joined: Fri Oct 29, 2010 1:37 am

Re: Major contact stability improvement

Post by VicariousEnt »

Erwin Coumans wrote:Thanks for sharing your ideas and work.

It will help if you update your work to the latest Bullet svn repository. I will create an up-to-date github repository, that makes it easier to fork/clone your modified version.
Erwin, what do you think of these changes? Do you agree with them? Do you see them causing issues with other areas that they may not have dealt with yet? And finally do you think these changes will make it into 2.81?

The game my company is working on is suffering from alot of jitter problems with ragdolls hanging, falling onto and being dragged over BVH meshes, these changes sound like they could fix alot of our problems. I would love to just wait for the 2.81 release though as we've been putting off upgrading to the latest version of bullet (we're on 2.78) untill the CCD system works better with ragdolls\constraints (which these changes could improve or fix by the sounds of it).
VicariousEnt
Posts: 50
Joined: Fri Oct 29, 2010 1:37 am

Re: Major contact stability improvement

Post by VicariousEnt »

Laurent Coulon wrote:Hi Erwin,
Thanks for the reply. I will try to make a patch for the latest bullet version. Let me know what the name of the repository is on github. Otherwise I can post it on Google code. Let me know.
Thanks for actually taking the time to post your discoveries! If these other changes and fixes you guys have made to Bullet or not specific to your game, please post them to the issue tracker as well. Would love to hear what else you have changed.
Laurent Coulon
Posts: 29
Joined: Tue Oct 05, 2010 9:36 pm

Re: Major contact stability improvement

Post by Laurent Coulon »

Hi Erwin,
I have grabbed the latest Bullet build. I can integrate my changes into it. Let me know how I should do this. Is there a specific repository I should use on Github? Or can I create my own svn branch on Google code?
Thanks for letting me know how to proceed.
trollington
Posts: 8
Joined: Thu Nov 24, 2011 3:08 pm

Re: Major contact stability improvement

Post by trollington »

Any news?
arno
Posts: 2
Joined: Sun Jul 01, 2012 11:11 am

Re: Major contact stability improvement

Post by arno »

Hello Laurent and Erwin,

This topic is going down in the list, so here is a little post to get it back on top.

I am very interested by the stability changes you described, Laurent. I have a small robot simulator, used to test things for the Eurobot contest.
This year I had troubles with some elements, probably due in part to size differences (some elements are only 2 mm thick, whereas the table is 3 meters long). The behavior I observed was that some elements that should be resting started to move randomly before stabilizing. I tested a lot of parameters, but it didn't removed all these stability problems.

It seems that your changes could help on these points. I tried to compile a version of bullet including the files you posted there. However there were many errors and missing parts to do it just like that. I managed to successfully compile it, but when I run it, nothing moves. I must have done something when trying to make it to compile, but I don't have a good background of the internals of bullet (and more generaly, I only know the basics of dynamics engine), so I'm stuck here.

I hope you could post a patch for the stability change, or upload it to any place (google-code, github, or any other...).
If it helps, I can then test it against my simulator, and provide feedback about the improvements.

Bullet is a very powerful and easy to use library. I hope these changes will get through it (or at least be made available), as it seems to be exactly what I need.

Thanks in advance on this point, and a more general thanks to all people making it possible for people like me to have access to such a powerful engine !

Arnaud
User avatar
Erwin Coumans
Site Admin
Posts: 4221
Joined: Sun Jun 26, 2005 6:43 pm
Location: California, USA
Contact:

Re: Major contact stability improvement

Post by Erwin Coumans »

@Laurent Coulon:
Please pick a recent trunk of Bullet, apply your changes, and provide a SVN patch and let use know what SVN revision you used.
You can upload a patch to https://code.google.com/p/bullet/issues/list

@arno Can you provide a reproduction case in any of the Bullet demos (Bullet/Demos/BasicDemo)?
Laurent Coulon
Posts: 29
Joined: Tue Oct 05, 2010 9:36 pm

Re: Major contact stability improvement

Post by Laurent Coulon »

Thanks Erwin, will do. I'll post again here when I have it done.
Post Reply